An Idea

Add structures like trees, pepole, houses, cars etc.(as stickers, so that they can easily show the AI where they really want the thing.) as special objects and let the users make their imagination images roughly on a blank page. After that, let the AI show them the enhanced version of the image they made. Show them multiple options for the same image. With this type feature, the people like us who don't know the English language well, will also be able to do the image generation easily. And also the other people don't have to give prompt again and again to make the AI understand. I am facing the same problem. That's how the idea came into my mind.